This week’s roundup covers capital strategy, Conway’s Law in the age of AI agents, LinkedIn ad benchmarks with real numbers, the Miro and Airtable exits, and a first-person account of what Bending Spoons does after it closes a deal.
Capital as a competitive moat
The Era of Compounding Capital argues that capital structure now functions as a competitive advantage, especially for AI and infrastructure companies. The core thesis: each funding milestone should reduce financial risk and attract more diverse funding sources. The most resilient companies use abundant capital to lock in cheaper, more durable capital next time, so every dollar raised makes the next raise easier.
Conway’s Law meets AI
This short post asks whether Conway’s Law still holds when AI is part your org’s communication structure. The answer: the law still applies, but the unit of communication has changed. Existing workflows need to shift toward outcomes rather than process. Human judgment still matters, which means humans have to adapt too, not just the tooling.
Paul Graham on startup power
Making Startups Powerful is a 15-minute read arguing founders should optimize for power, not just profit. The strategies named: owning customer relationships, building network effects, becoming a marketplace or full-stack solution, helping users make money, and selling early-stage to customers who decide fast.
LinkedIn ads: the real benchmarks
This LinkedIn ad playbook opens with a number that stops most solo operators cold: $37.42 CPM versus roughly $10 on Meta. The author’s criteria for when LinkedIn ads actually make sense: deal size of at least $10K, a minimum of 10,000 reachable decision-makers, and a test budget of at least $3,000 a month. The targeting approach skips job titles entirely and instead pulls buying signals like job postings and tech-stack changes, uploaded as custom audiences. Switching one account to manual bidding dropped cost per lead from $287 to $90.

Miro and Airtable: cash-flow positive, sold at 2.5x revenue
This SaaStr post is a cold bucket of water on the cash-flow-positive narrative. Miro had roughly $600M in ARR and sold for $1.355B. Airtable had roughly $480M in ARR, growing over 20%, and sold for $1.285B. Together they held about $1.4B in cash, and buyers paid nothing extra for it. Growth rate was the only thing priced. The cash earned the worst multiple on the cap table.
What Bending Spoons looks like from the inside
An engineer who survived the Evernote acquisition wrote up what the six months after close actually looked like. Evernote was around $100M in ARR when 129 people were laid off in a single round. His team went from six developers to one. The annual subscription price moved from $69.99 to $129.99. Bending Spoons later listed at $18.4B. The writer’s read on a voluntary severance offer: it’s the new owner announcing its plan for the org. Relevant context given that Bending Spoons also acquired Miro and Airtable this month.
⚡ Quick links
- PostHog scaled from 1 to 100+ in-person events a year by removing the approval loop. Engineers demo the features they built, travel budget is pre-provisioned, and no one needs sign-off. Over half the company has demoed somewhere. About 95% of events put them face-to-face with customers. 20% of engineers opt out and nothing happens to them.
- Slack Surfaces turns conversations into live dashboards, decks, or reports that update automatically from their source and can be pinned into a channel for the team to filter, comment on, and act from together.
- LangChain’s paid media agent teardown: the strongest results came from giving the agent a sandbox, business context, and clear operating instructions. Use models for judgment, code for consistency. The agent proposes campaign changes, routes them through human approval, then verifies they were applied correctly.
- PLG-to-enterprise sales pain points are completely unoriginal, according to Elena Verna. Every PLG company hits the same walls. The packaging differs, the conversations don’t.
